gstlal.far: move .add_background_prior(), .add_foreground_prior()

from DistributionStats class to CoincParamsDistributions class

gstlal-inspiral/bin/gstlal_inspiral_calc_likelihood

@@ -145,11 +145,11 @@ coincparamsdistributions, likelihood_seglists = Far.distribution_stats, Far.live
 
 # calculate injections before writing to disk
 if options.synthesize_injections != 0:
-	coincparamsdistributions.add_foreground_prior(n = options.synthesize_injections, instruments = Far.trials_table.get_sngl_ifos(), verbose = options.verbose)
+	coincparamsdistributions.raw_distributions.add_foreground_prior(n = options.synthesize_injections, instruments = Far.trials_table.get_sngl_ifos(), verbose = options.verbose)
 
 # add a uniform prior to background, by default 0 is added so it has no effect
 if options.background_prior != 0:
-	coincparamsdistributions.add_background_prior(n = options.background_prior, instruments = Far.trials_table.get_sngl_ifos(), verbose = options.verbose)
+	coincparamsdistributions.raw_distributions.add_background_prior(n = options.background_prior, instruments = Far.trials_table.get_sngl_ifos(), verbose = options.verbose)
 
 if options.verbose:
 	print >>sys.stderr, "computing event densities ..."

gstlal-inspiral/bin/gstlal_inspiral_create_prior_diststats

@@ -100,9 +100,9 @@ options, filenames = parse_command_line()
 coincparamsdistributions = far.DistributionsStats()
 
 if options.background_prior != 0:
-	coincparamsdistributions.add_background_prior(options.background_prior, instruments = options.instrument, verbose = options.verbose)
+	coincparamsdistributions.raw_distributions.add_background_prior(options.background_prior, instruments = options.instrument, verbose = options.verbose)
 
-coincparamsdistributions.add_foreground_prior(verbose = options.verbose, n = options.synthesize_injection_count, instruments = options.instrument)
+coincparamsdistributions.raw_distributions.add_foreground_prior(verbose = options.verbose, n = options.synthesize_injection_count, instruments = options.instrument)
 
 # no segments, this file is independent of time
 FAR = far.LocalRankingData(segments.segment(None, None), coincparamsdistributions)

gstlal-inspiral/python/far.py

@@ -417,35 +417,8 @@ class ThincaCoincParamsDistributions(snglcoinc.CoincParamsDistributions):
 			instruments.discard("H2")
 		return dict(("%s_snr_chi" % event.ifo, (event.snr, event.chisq / event.snr**2)) for event in events if event.ifo in instruments)
 
-
-#
-# Paramter Distributions
-#
-
-
-class DistributionsStats(object):
-	"""
-	A class used to populate a ThincaCoincParamsDistribution instance using
-	event parameter data.
-	"""
-	def __init__(self):
-		self.raw_distributions = ThincaCoincParamsDistributions()
-		self.smoothed_distributions = ThincaCoincParamsDistributions()
-		self.likelihood_pdfs = {}
-		self.target_length = 1000
-
-	def __add__(self, other):
-		out = type(self)()
-		out.raw_distributions += self.raw_distributions
-		out.raw_distributions += other.raw_distributions
-		#FIXME do we also add the smoothed distributions??
-		return out
-
-	def add_single(self, event):
-		self.raw_distributions.add_background(self.raw_distributions.coinc_params((event,), None))
-
 	def add_background_prior(self, n = 1., transition = 10., instruments = None, prefactors_range = (1.0, 10.0), df = 40, verbose = False):
-		for param, binarr in self.raw_distributions.background_rates.items():
+		for param, binarr in self.background_rates.items():
 			new_binarr = rate.BinnedArray(binarr.bins)
 			# FIXME only works if there is a 1-1 relationship between params and instruments
 			instrument = param.split("_")[0]
@@ -469,51 +442,14 @@ class DistributionsStats(object):
 			# add to raw counts
 			binarr += new_binarr
 
-		# FIXME, an adhoc way of adding glitches, us a signal distribution with bad matches
-		# FIXME if we keep this, make a function that can be reused for both signal and background
-		# FIXME:  for maintainability, this should be modified to
-		# use the .add_injection() method of the .raw_distributions
-		# attribute, but that will slow this down
-		pfs = numpy.linspace(prefactors_range[0], prefactors_range[1], 10)
-		for param, binarr in self.raw_distributions.background_rates.items():
-			new_binarr = rate.BinnedArray(binarr.bins)
-			# FIXME only works if there is a 1-1 relationship between params and instruments
-			instrument = param.split("_")[0]
-			# save some computation if we only requested certain instruments
-			if instruments is not None and instrument not in instruments:
-				continue
-			if verbose:
-				print >> sys.stderr, "synthesizing background for %s" % param
-			# Custom handle the first and last over flow bins
-			snrs = new_binarr.bins[0].centres()
-			snrs[0] = snrs[1] * .9
-			snrs[-1] = snrs[-2] * 1.1
-			chi2_over_snr2s = new_binarr.bins[1].centres()
-			chi2_over_snr2s[0] = chi2_over_snr2s[1] * .9
-			chi2_over_snr2s[-1] = chi2_over_snr2s[-2] * 1.1
-			for i, snr in enumerate(snrs):
-				for j, chi2_over_snr2 in enumerate(chi2_over_snr2s):
-					chisq = chi2_over_snr2 * snr**2 * df # We record the reduced chi2
-					dist = 0
-					for pf in pfs:
-						nc = pf * snr**2
-						v = stats.ncx2.pdf(chisq, df, nc)
-						if numpy.isfinite(v):
-							dist += v
-					dist *= (snr / snrs[0])**-4
-					if numpy.isfinite(dist):
-						new_binarr[snr, chi2_over_snr2] += dist
-			# normalize to the requested count
-			new_binarr.array *= n / new_binarr.array.sum()
-			# add to raw counts
-			binarr += new_binarr
+		# FIXME, an adhoc way of adding glitches, use a signal distribution with bad matches
+		self.add_foreground_prior(n = n, prefactors_range = prefactors_range, df = df, instruments = instruments, verbose = verbose)
 
 	def add_foreground_prior(self, n = 1., prefactors_range = (0.0, 0.10), df = 40, instruments = None, verbose = False):
 		# FIXME:  for maintainability, this should be modified to
-		# use the .add_injection() method of the .raw_distributions
-		# attribute, but that will slow this down
+		# use the .add_injection() method of self, but that will slow this down
 		pfs = numpy.linspace(prefactors_range[0], prefactors_range[1], 10)
-		for param, binarr in self.raw_distributions.injection_rates.items():
+		for param, binarr in self.injection_rates.items():
 			new_binarr = rate.BinnedArray(binarr.bins)
 			# FIXME only works if there is a 1-1 relationship between params and instruments
 			instrument = param.split("_")[0]
@@ -521,7 +457,7 @@ class DistributionsStats(object):
 			if instruments is not None and instrument not in instruments:
 				continue
 			if verbose:
-				print >> sys.stderr, "synthesizing injections for %s" % param
+				print >> sys.stderr, "synthesizing background/injections for %s" % param
 			# Custom handle the first and last over flow bins
 			snrs = new_binarr.bins[0].centres()
 			snrs[0] = snrs[1] * .9
@@ -546,6 +482,33 @@ class DistributionsStats(object):
 			# add to raw counts
 			binarr += new_binarr
 
+
+#
+# Paramter Distributions
+#
+
+
+class DistributionsStats(object):
+	"""
+	A class used to populate a ThincaCoincParamsDistribution instance using
+	event parameter data.
+	"""
+	def __init__(self):
+		self.raw_distributions = ThincaCoincParamsDistributions()
+		self.smoothed_distributions = ThincaCoincParamsDistributions()
+		self.likelihood_pdfs = {}
+		self.target_length = 1000
+
+	def __add__(self, other):
+		out = type(self)()
+		out.raw_distributions += self.raw_distributions
+		out.raw_distributions += other.raw_distributions
+		#FIXME do we also add the smoothed distributions??
+		return out
+
+	def add_single(self, event):
+		self.raw_distributions.add_background(self.raw_distributions.coinc_params((event,), None))
+
 	def finish(self, verbose = False):
 		self.smoothed_distributions = self.raw_distributions.copy()
 		#self.smoothed_distributions.finish(verbose = verbose)